Rakuten, Inc.

Masato Hagiwara

Soh Masuko

Ayah Zirikly

A multilingual named-entity recognition system according to an embodiment includes an acquisition unit configured to acquire an annotated sample of a source language and a sample of a target language, a first generation unit configured to generate an annotated named-entity recognition model of the source language by applying Conditional Random Field sequence labeling to the annotated sample of the source language and obtaining an optimum weight for each annotated named entity of the source language, a calculation unit configured to calculate similarity between the annotated sample of the source language and the sample of the target language, and a second generation unit configured to generate a named-entity recognition model of the target language based on the annotated named-entity recognition model of the source language and the similarity.

Computer system, method and program for performing multilingual named entity recognition model transfer

To allow for easy entry of a plurality of characters by handwriting gestures in the air, a user interface device includes template data storage means for storing template data indicating changes in a predetermined writing position when a gesture to write each of a plurality of characters in the air is made, position obtaining means for sequentially obtaining the predetermined writing position when a user makes gestures to sequentially writing characters in the air, similarity evaluation information output means, every time the predetermined writing position is obtained by the position obtaining means, for sequentially outputting similarity evaluation information indicating a similarity between data to be evaluated including a predetermined number of the predetermined writing positions taken in order from newly obtained data and the template data related to each of the plurality of characters, and character string determination means for determining a character string related to the gestures of the user based on the sequentially output similarity evaluation information related to each of the plurality of characters.

User interface device, user interface method, program, and computer-readable information storage medium

To enable a user to easily select a selection element, a user interface device includes image capturing means for capturing an image of a user viewing a display screen, recognition means for recognizing the position of a reference part of the user's body appearing in the captured image, which differs from a selection position that is a position for the user to select a selection element displayed on the display screen, display position determination means for determining a display position of the selection element based on the position of the reference part appearing in the image, and display means for displaying the selection element in the determined display position.

A transliteration device () comprises a generation part () generating rewriting tables corresponding to K different languages and including multiple rewriting probabilities that an original segment can be rewritten as a transliterated segment for transliteration and transliteration tables corresponding to the K languages. The transliteration device () further comprises an update part () saving the transliteration probability that an original spelling string originating from a language corresponding to the used rewritten language is transliterated to a target spelling string in the transliteration able corresponding to the language, and so updating the K rewriting tables as to maximize the expected value of a likelihood function calculating the likelihood of the K transliteration tables.

Transliteration device, transliteration program, computer-readable recording medium on which transliteration program is recorded, and transliteration

A receiving unit () receives a seed string. A search unit () searches snippets of documents containing the seed string. A segment acquisition unit () obtains segments by partitioning the snippets using a segment partition string. A segment component acquisition unit () obtains segment components by partitioning the segments using a segment component partition string. A segment score computation unit () calculates a segment score for a segment based on the standard deviation of the lengths of the segment components. A segment component score computation unit () calculates a segment component score for a segment component based on the segment score and the distance between the position of the seed string and the position of the segment component. A selection unit () selects any of the segment components as candidates for instances contained in the expanded set of the seed string based on the segment component scores.

Growing community of inventors

Tokyo, Japan

Masato Hagiwara